What field-service software actually charges in 2026 — a buyer's guide for 1–10 person trade businesses.
Every trade app now says "AI." Almost none of them make it easy to answer the only two questions that matter: what does the AI actually do, and what does it actually cost by the time it's doing it for your business? The pricing pages are ladders, the AI features are scattered across tiers and add-ons, and half the vendors won't publish a number without a sales call.
We sell software in this market, so read this knowing that. But every number below is sourced from the vendor's own website or help docs, dated, and checkable in five minutes. Where a fact favors a competitor, we've printed it anyway — there's a section on what they include that we don't.
When a trade app says AI, it means one of two very different things:
Answers questions and takes messages. It's a receptionist, a report-writer, a tour guide. This is real technology — some of it genuinely sophisticated. But notice what's still true after you buy it: the quoting, invoicing, follow-ups, and books are exactly as manual as before. Someone still does the office work. The AI just talks about it.
Does the office work. It drafts the invoice when the job's marked done. It notices the quote that's gone quiet for five days and writes the follow-up. It watches the overdue invoice and prepares the reminder — tone matched to how late it is. It closes the books at night. A human approves; the AI executes.
Almost everything sold as "AI" in this market in 2026 is the first kind. The comparison tables below are about what the first kind costs — and what it can't do at any price.
Here's a number nobody publishes: how many AI actions does one working tradesperson's month actually take? From real usage on our platform (June 2026, one active solo operator):
Hold that number. It's the yardstick for every "AI included" claim below — because most of those claims come with a meter.
| Vendor | Entry price | What a 1–5 person shop really pays | Where the AI lives |
|---|---|---|---|
| Jobber | $39/mo Core (1 user) | Connect Team $169 (5 users) → Grow Team $349 (10 users) → Plus $599 (15 users) | AI Receptionist: $29/mo for 30 conversations, $0.79 each after — unlimited only on Plus, $599/mo |
| Housecall Pro | $79/mo Basic (1 user; $59 annual) | Essentials $189/mo (5 users; $149 annual) → MAX $329/mo + $35/user ($299 annual) | Analyst/Coach/chat AI included on all plans (genuinely). 24/7 AI call answering: paid add-on, contact sales. QuickBooks sync: gated at Essentials |
| ServiceTitan | Unpublished — sales demo required | $250–500/technician/mo (user-reported) | Atlas assistant + Dispatch Pro exist and are capable — as paid Pro add-ons on unpublished pricing |
| Workiz | $225/mo Kickstart (3 users) | Standard/Pro $275–325/mo (5 users — both tiers cap at 5) | Genius AI needs Pro; Answering is a further ~$200/mo add-on (unpublished rate, user-reported) on a required ~$100/mo phone add-on — 5-seat, AI-enabled runs ~$625/mo total. No fixed contract found; ~17% off for annual prepay |
| QuoteIQ | $29.99/mo (1 user) | $75–150/mo (Beginner/Pro) for a working solo | Metered "IQ Credits": 500/mo at entry, no rollover. Their AI phone agent burns 125 credits/minute — the entry plan holds 4 minutes a month. Every AI action shows its price and asks you to confirm before running |
| Werkd | $99/mo flat (+$59/seat), or $999/yr (+$599/yr/seat) | $99 — everything below included, unmetered | 12 AI systems running the office work; no long-term contract |
No amount of spending buys it. Try to buy every capability in this guide at Jobber, using their published prices: unlimited AI reception requires Plus — $599/month. At Housecall Pro the published floor is MAX at $329/mo plus add-ons they won't price publicly (which tells you something by itself). ServiceTitan won't publish a starting number at all. And after all that spend, the middle rows of the worksheet in section 06 are still empty: the invoice still doesn't draft itself, the follow-up still doesn't write itself, the books still don't close at night. Not because you didn't pay enough — none of them sell it, at any price or any tier. It doesn't exist in their product.
Metered AI has two costs. The dollars are the obvious one: 900 real actions a month doesn't fit in a 500-credit allowance, and per-conversation or per-minute overage math turns your software bill into a utility bill you can't predict.
The second cost is the one that changes how you work: when every AI action shows you a price and asks permission to spend, you start rationing the help. You skip generating the estimate because the credits are already thin this month. You answer the phone yourself on the 25th because the credits are low. The thing you bought to reduce decisions is now generating a new decision, thirty times a day.
Take a modest reference month — roughly 25 jobs, 20 quotes, a few dozen customer conversations. (That's the workload behind our ~900-action telemetry in section 02. Busier than that? The meters get worse; a flat price doesn't.)
The cleanest published numbers in this market happen to be for phone/text reception — a category we don't build ourselves (see the candor note in section 05). We're printing them anyway: it's the rare case where a vendor's per-action math is fully public, and the same rationing logic applies to every metered AI feature — including the ones we do build, below.
Jobber's receptionist meter — fully computable from their published pricing ($29/mo for 30 conversations, $0.79 each after). This is the add-on cost alone, stacked on top of whatever Jobber plan you're already paying for — not the whole bill:
| Conversations in a month | Receptionist add-on (alone) | Total with Jobber Core ($39/mo) |
|---|---|---|
| 30 | $29.00 | $68.00 |
| 60 | $52.70 | $91.70 |
| 100 | $84.30 | $123.30 |
| 150 | $123.80 | $162.80 |
| No meter at all | Plus plan — $599/mo total (replaces Core, not stacked on it) | |
So a realistic month — 60 conversations — runs a Jobber Core customer $91.70 total (base plan plus the reception add-on), and that spend buys exactly one thing: automated reception. It doesn't touch invoicing, quote follow-ups, book-closing, or any of the other systems in section 05. Werkd is $99 flat — $7.30 more than that Jobber total — and includes all twelve systems, not just reception.
QuoteIQ's phone meter — computable from their one published rate (125 credits/minute): fifteen answered calls at ~90 seconds each is roughly 2,800 credits — past the $29.99 tier (500 credits) and the $74.99 tier (1,500) before a single estimate, image, or CoPilot edit has run. Call answering alone forces the $149.99 tier. Adjust the call count to your month; the rate is theirs.
Everything else on QuoteIQ's meter — we tried, and couldn't. Estimates, before/after images, and CoPilot edits each "consume more per action" (their words), but the per-action rates aren't published anywhere we could find. We set out to calculate what a normal month of AI estimating costs on IQ Credits and could not price it in advance. A meter you can't price before you're on it is the meter problem in one sentence.
Where it's genuinely apples-to-apples: photo diagnosis. We build this — snap a photo, get a priced draft, unmetered, run it as many times as the job requires. QuoteIQ builds a version of it too, under "estimates" and "before/after images" — drawing from that same unpublished-rate credit pool. Jobber's phone meter above is computable because they publish the rate. The one competitor feature that goes head-to-head with something we actually built isn't computable at all.
Werkd is $99/month, flat, month-to-month, plus $59 per additional seat — or billed annually at $999/year for the owner and $599/year per additional seat, saving $189 and $109/seat respectively. Annual isn't a lock-in either: cancel any time and unused months are refunded. That price includes 12 AI systems that do office work, unmetered: photo diagnosis (what's actually wrong, priced from real parts and labor — not a square-footage guess), a composer that turns a sentence into a scheduled job or quote, and a receipt photo into a logged expense with vendor, tax, and line items already read off it, nightly book-closing, payment chasing with tone matched to lateness, a morning desk of prepared decisions, a money brief that names the one thing most worth chasing today, a price book that learns your real costs from receipts, route and time-window checking, seasonal and win-back nudges, pre-arrival job briefings, cross-sell suggestions, and review requests timed per customer. Beyond those twelve, smart booking offers customers real open times to choose from — scored by drive time, existing jobs, and work hours, not a raw calendar grid — and every Monday, a written business report (wins, concerns, what to watch) is drafted from that week's real numbers and sent automatically.
And the part that makes working AI trustworthy: nothing customer-facing sends itself blindly. Drafts are shown before they go. Reminders run on the schedule you set, and any customer can be switched to manual review — the AI keeps drafting, you approve every send. Every action lands in a full audit log: what ran, when, why.
Be conservative with our own number: say only a third of those ~900 monthly actions would otherwise be done by you, at five minutes each. That's ~25 hours a month of office work — done after dinner, because that's when office work happens for a working tradesperson. Price it two ways: at your own billable rate (25 × $___/hr = $___), or against the alternative — a part-time office admin runs $1,500–3,000/month. Werkd is $99. Cut our assumptions in half if you like; the math survives. You do the arithmetic — we just did the drafting.
Unmetered means no credits, no counters, no per-action prices, and no overage bill — ever, for running your business. Your busiest month — the storm week with 200 photos and a backlog of invoices — never sees a meter. What "unmetered" doesn't mean is infinite compute for things that aren't running a trade business — scripted bulk traffic, reselling access, running three companies through one login. Like every flat-price service, our terms carry a fair-use line for those cases. If you're a tradesperson doing tradesperson work, you will never meet it.
So you hear it from us: no AI answering customer texts or calls — when a customer reaches out, it comes straight to you, every time; texts to customers send from your phone with one tap (we're deliberately not the sender of record — no Twilio account, no 10DLC registration, no per-message fee, but it does mean your thumb hits send); and no enterprise dispatch board — we're built for 1–10 person operations, not 50-truck fleets. If you need those, some of the vendors above are the right call.
Take this to any demo. For each row, ask: included, add-on, or not available? — and get the monthly number in writing.
| The office job | Werkd | Ask your vendor |
|---|---|---|
| AI answers customer messages | Not offered — goes to you directly | Which tier? Metered? Overage rate? |
| Review requests, timed & drafted | Included | Add-on price? |
| Quote follow-ups written for you | Included | Does it draft, or just remind you? |
| Invoices drafted when the job's done | Included | Does this exist at any tier? |
| Payment reminders, tone-matched, on schedule | Included | Auto or manual? Per-account control? |
| Books closed nightly with P&L | Included | Does this exist at any tier? |
| Your real parts costs learned from receipts | Included | Does this exist at any tier? |
| A full log of everything AI did | Included | Can you see what the AI actually did? |
| Total monthly, all rows, your team size | $99 + $59/seat | $_________________ |
The pattern you'll find: the first two rows are purchasable everywhere (as add-ons or tiers). The middle rows — the working AI — mostly aren't available at any price. That's not a pricing gap. It's an architecture gap: software built as a system of record can bolt an assistant onto the front desk, but it can't retrofit an employee.
14-day trial, no sales call required — self-serve is the point.
Prices are monthly-billing (no-commitment) rates from vendor websites and help documentation, captured July 3–5, 2026. Annual billing reduces most of them, most dramatically at Jobber (up to ~40%) and Workiz (~17%). Usage telemetry is from Werkd production logs, June 2026, one active solo account. Vendors change pricing frequently — if any number above is out of date, email us and we'll correct it in the next revision.
Werkd is built for independent tradespeople and crews of 1–10. Everything in this guide about our own product is verifiable in a 14-day trial.